Mario Rossi

Dr. Mario Rossi


Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
Verified
Journal Article
Open access
PDF file
DGDaniel Gattari
Daniel Gattari
JSJoseba Sancho-Zamora
Joseba Sancho-Zamora
DCDebora Mirta Chan
Debora Mirta Chan
NPNatalia Jorgelina Prado
Natalia Jorgelina Prado
EDEmiliano Diez
Emiliano Diez
MRMario Rossi
Mario Rossi
Connexin-43 (CX43) lateralization in ventricular myocardium has been associated with abnormal impulse propagation and increased arrhythmia susceptibility. Its quantitative assessment in histological sections remains challenging because previous methods require segmentation of individual cardiomyocytes and rely on geometric rules applied to segmented cell profiles. Here, we present CLARISA, a segmentation-free, ROI-based deep learning framework that classifies CX43-positive regions as terminal or lateralized directly from fluorescence images. An expert-annotated dataset was generated from left-ventricular cryosections of Wistar rat hearts, in which CX43-positive regions were labeled according to their distribution pattern. A dual-stream EfficientNetV2-S classifier was trained to capture both local and contextual ROI morphology. We also developed a semi-automated whole-section inference module to generate spatial lateralization probability maps and global percent lateralization estimates. On the held-out test set, CLARISA achieved a ROC-AUC of 0.904 (95% bootstrap CI: 0.828–0.960) and a PR-AUC of 0.808 (95% bootstrap CI: 0.682–0.913), supporting the feasibility of automated ROI classification for CX43 lateralization assessment. When deployed on whole tissue sections, including an independently analyzed section not used during model development, CLARISA generated spatial maps that captured heterogeneous CX43 organization and produced a global percent lateralization estimate closely aligned with expert annotation, differing by only 1.30 percentage points over the same detected CX43-positive area. Comparison with a previously published segmentation-based method further indicated that ROI-based and cell-segmentation-based approaches provide related but non-equivalent readouts of CX43 lateralization. The ROI-based design additionally reduces annotation burden—requiring classification of discrete CX43-positive signal rather than complex cardiomyocyte delineation—and ensures that all detected CX43-positive signal contributes to the lateralization estimate regardless of cell boundaries. These results establish CLARISA as a proof-of-principle framework for scalable, segmentation-free CX43 lateralization assessment in cardiac tissue. Further validation across larger, independent, and more heterogeneous datasets will be required to assess robustness, portability across imaging conditions, and translational applicability. The complete codebase, pretrained model, image data, and expert annotation tool are publicly available.

2 May 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
DGDaniel Gattari
Daniel Gattari
JSJoseba Sancho-Zamora
Joseba Sancho-Zamora
DCDebora Mirta Chan
Debora Mirta Chan
EDEmiliano Diez
Emiliano Diez
MSMariano Llamedo Soria
Mariano Llamedo Soria
MRMario Rossi
Mario Rossi
Connexin-43 (CX43) lateralization in ventricular myocardium has been associated with abnormal impulse propagation and increased arrhythmia susceptibility. Its quantitative assessment in histological sections remains challenging because of the difficulty of segmenting individual cardiomyocytes and the reliance of previous methods on geometric rules applied to segmented cell profiles. Here, we present CLARISA, a deep learning framework for classifying CX43-positive regions as either terminal or lateralized directly from fluorescence images, without requiring cardiomyocyte segmentation. An expert-annotated dataset was generated from left-ventricular cryosections of Wistar rat hearts, in which CX43-positive regions were labeled according to their distribution pattern. A dual-stream convolutional classifier based on EfficientNetV2-S was trained to capture both the local and contextual morphology of each region. In addition, an inference module applicable to whole tissue sections was developed to generate spatial lateralization probability maps and global percent lateralization estimates consistent with expert annotation. On the test set, CLARISA achieved a ROC-AUC of 0.905 and a PR-AUC of 0.810. These results support the feasibility of automated assessment of CX43 distribution patterns without explicit cardiomyocyte segmentation. The complete codebase is publicly available, together with access to the pretrained model and the image data used in this study. The Hugging Face model card reports the same held-out test metrics and states that the checkpoint is intended to be used with the main repository.